increasing anchor embedment depth in an existing slab

huangyhg

increasing anchor embedment depth in an existing slab
hi all
i have a tank structure (with four legs) that needs to be anchored to an existing concrete slab. the chemically fixed anchors i want to use require a min embedment depth of 300mm. i can only get 200mm embedment out of the existing slab.
i can pour local footings/plinths on top of the existing slab, one under each leg, to increase my embedment depth. i can make these large enough, in plan, to accommodate the theoretical projected area of a cone pull out failure.
question is, if i simply scabble the existing surface and pour a new concrete footing on top, will the two act as one (ie as if it had all been poured at the same time) in terms of a cone pull out failure? or should i add extra steel reinforcement to "dowel" the new slab into the existing to prevent the top one trying to prematurely lift away from the bottom in the even of cone out type failure.
the extra plinths im considering would have dimensions of 1500mm x 1500mm x 100mm thk.
appreciate any thoughts
thanks

check out our whitepaper library.
don't agree here as the slab is what you have to anchor to, unless you provide the same capacity in the attachment of the plinth to the slab as the overall anchor bolt capacity you are trying to generate.
why not sawcut a square hole in the slab and overexcavate under the slab, generating enough beam shear to resist the uplift.  the new footing could be made deep enough to provice the required embedment of the anchor bolts.
mike mccann
mmc engineering
in cone pull failure even if bond along the shaft takes part on forming the conical struts passing shear to the external one considered for failure, it is safer to assume that, in lack of any bond, the whole tensile load is passed in keying action to the end of the deep end of the anchor. it is obvious this can only be a notional artifac for simply adhered anchors with no particular deep end fixity or mechanical action, yet it might make sense from a design wanting to be conservative; in this light, the advice given by msquared48 of passing the whole load to the existing slab would be adequate.
if, on the other hand, out of the nature of the continuos adhesion to the shaft one assumes a uniform distribution (or whichever reasonable) one just would have to pass to the existing slab the part tributary to its embedded length ... with a particular precision, that, as you note, the integrity between the existing and new concrete be enough to preserve the cone failure solidary behaviour.
one can imagine, as you have done, the added concrete part, separating with the plate outfit from the existing slab. opposing this separation are the tying action of the adhered anchor and the tensile strength adhesion at the interface, taken only at the design value at the circle intersection with the cone at the interface (this can be made however very high -even with prefabricated plinths- by using methyl-methacrylate bonding agents or epoxy on some clean interface).
then, in akin way of thinking to when proposing to pass the whole tensile action to the existing slab, one safe approach would be to dismiss altogether any tying action by the anchor, and let the adhesion between old and new concrete -more clearly without any bonding agents, that would introduce a perturbation on the cone formation- deal just with the tensile action considered to be being passed to the existing slab, that at a safe design value.
then, as well, and since your plinths are of finite size, you may as well model some trio of rectangular prisms for plate, plinth and some extent of slab (constraints to deal with equilibrium against tumbling), join,  and applying the tensile action at the plate or some bolt holes, derive some horizontal shear stresses appearing at the plinth-existing slab interface.  
then you can reinforce in shear friction for the stresses there, assuming an old-new interface friction factor that i re  
motorspirit...chemical anchors don't fail in typical tensile-shear cone failure.  they fail in circumferential shear on the face of the drilled hole.
why are your uplift loads high?  overturning?  do you have enough footing mass to take care of overturning?
dowel the new concrete into the old and use an epoxy bonding agent.  that should make them act as monolithic.
mike's idea of cutting through the existing slab and extending the footing downward is good.  we use this method to anchor aluminum canopies.  it gives us more resisting mass that is engaged by the shear key at the bottom of the slab.
mike's idea is indeed good, but i expect most contractor's would carry out any amount of dowel's to ensure a plinth cast on top of the slab was monolithic rather than digging down.
but, what loads are we talking about? is the slab designed for these loads or is mike's idea required anyway to form a foundation?
if the loads aren't significant it may be that the 300mm embedment is just a guidline from the epoxy supplier and a designed anchor may be possible with 200mm anchorage length.
to add that apart from the shear stresses there would also be direct tensile stresses in the case one goes for a 3d model, and the interaction of both would gover pinning desingn if needed.
thanks for your input guys.
the existing slab forms part of a secondary containment bund and therefore was looking to avoid cutting into it, otherwise i would definitely have gone with mikes solution. i can still do it this way but would have to be very sure i seal it up again properly.
i agree that if i place a plinth on top i would need to anchor it to the existing slab with doweling of equivalent capacity.
ron, im not sure i follow your first point. all of the info i have from chem anchor suppliers suggests that if concrete failure is the critical case, it will be a cone failure??
fyi the uplift is generated due to seismic loading. the max uplift to be resisted by the anchor group on each leg is 140kn. the anchorage needs to meet capacity design principles, ie the anchors must reach overstrength yield before concrete failure.
im still not sure which way ill end up going with this one but im erring on the side that mike suggests
thanks again
how about modifying the base plates to get more anchor rods to take part?  maybe a thicker plate with buttresses and 6 or even 8 rods each corner.  yes, they lose individual capacity when put close together, but you should be able to arrange them such that they'll have a greater total capacity...
if you "heard" it on the internet, it's guilty until proven innocent. - dcs
motorspirit...the epoxy anchors depend on their bond to the inside face of the drilled hole to mobilize stress in the concrete.  due to variations in the techniques of drilling and cleaning the hole, this is often the "weak link" in the process.
if you compute the area of the hole face and apply a shear-bond strength of say 200 to 300 psi, then compare a cone area and the tensile strength of the concrete (8 to 10% of f'c), you'll find it takes a lot more force to mobilize the tensile stress in the cone than to fail the hole interface in shear.
several manufacturers have tested this and will likely provide test data (i know that simpson did this for their anchors....we had them test some anchors on a jobsite once for this condition...they failed in shear, not in tension over cone area).